This function is powerful. You may:
- search in any column of range (in data base);
- define index number in search result.

Look at attached XLS-file for examples.

But there is a room for improvement (original function have slow algorithm)

Function VLOOKUP2(Table As Variant, SearchColumnNum As Long, SearchValue As Variant, _ N As Long, ResultColumnNum As Long) Dim i As Long, iCount As Long Select Case TypeName(Table) Case "Range" For i = 1 To Table.Rows.Count If Table.Cells(i, SearchColumnNum) = SearchValue Then iCount = iCount + 1 End If If iCount = N Then VLOOKUP2 = Table.Cells(i, ResultColumnNum) Exit For End If Next i Case "Variant()" For i = 1 To UBound(Table) If Table(i, 1) = SearchValue Then iCount = iCount + 1 If iCount = N Then VLOOKUP2 = Table(i, ResultColumnNum) Exit For End If Next i End SelectEnd Function

Tasks for optimization:
- change from loop to .find (but I not know .find syntax);
- optimize search algorithm if user want to retrieve only first value in records set;
- optimize search algorithm if user want to retrieve not first value in records set;
- optimize search algorithm if user want to retrieve more as one value in records set;
- add option for return count of records set.

This uses Array method rather than cell processing (so will be far faster)
Test it, see if it does the job.

you would not use a .find type as it will be slower than just looping an array.
Setting up and running a find would be slower than looping an in memory array.

Function VLOOKUP3(Table As Range, SearchColumnNum As Long, SearchValue As Variant, N As Long, ResultColumnNum As Long) Dim i As Long, iCount As Long Dim DataArr DataArr = Table.Value For i = 1 To UBound(DataArr) If DataArr(i, SearchColumnNum) = SearchValue Then iCount = iCount + 1 End If If iCount = N Then VLOOKUP3 = DataArr(i, ResultColumnNum) Exit For End If Next iEnd Function

This will speed up the existing function, but not really change it.
I am sure it is possible to do each of the changes you require, but I would suggest a single step at a time witha single question for each.

To cover all the steps in one hit is a bit more than a 300 point question :)

The attached has four methods...
- VLookup2 - the original method, searching cell by cell.
- VLookup3 - The_Barman's array method.
- VLookup4 - uses Find().
- VLookup5 - uses Match().

I don't believe that any single method is the optimum for all conditions. With the exception of the Find() method, each of the method is the best under some circumstances...
- The Cell method is best for a large search space, looking for a single value - which is found very early. Under these circumstances, the Array method is by a long way the worst. (Which makes sense - The Cell method has no overhead, but is expensive per row, whereas the Array method has a huge overhead, but once the array is loaded searching is effectively instant.)
- The Array method is best when searching a small number of cells - certainly less than 50. (At the other extreme, when looking for multiple hits, it's somewhere after the 4,500th occurrence that the advantage swings back the Array method.)
- Other than the above (special?) cases, the Match method is pretty much always the fastest method.
- There are various circumstances under which the Find method can beat the Cell or Array methods. However, a fundamental problem with Find is that you have to make a choice between searching by Values or Formulas. Values seems like the obvious choice, but that means that 12497 won't find that number if its formatted (e.g.$12,497.00) whereas searching by Formulas will. However, it would seem to be an unacceptable constraint that you couldn't search cells containing formulas, so Find is out.
- The Cell and Array methods are case-sensitve, whereas the Find and Match methods aren't. Potentially a show-stopper.

So...
- For case-sensitive, the Array method is best.
- For a small table, the Array method is best.
- For huge extremes of occurrence (e.g. the 5,000th occurrence of a value), the Array method is also best.
- For everything else, Match is best. It has the further advantage that I never saw it take a long time - the worst time was looking for the 8,000th occurrence of a value in 60,000 row which took a tenth of a second. At its best, it was 50 or more times faster than the Array method.

Before I started this, I assumed that the Array method would be the optimum approach, so Match's results were a big surprise to me. A word of warning though, the Array method is a common and well-tested method, whereas I've never seen the Match method used before.

Finally, points are officially a measure of the importance of the question to you, not the work content, so the 300 isn't an issue. However, I agree with The_Barman that you really have multiple questions here. I suggest that you treat this question as a performance one.

Once you've got the performance issues clear in your own mind (and unless performance is hugely significant, I'd go for the Array method) then open a second question for all the other functionality as a single function - it's easy to use optional parameters to include or exclude functionality for individual runs.

I hadn't considered the MATCH function as I would have expected it to be in line with the FIND method and ultimatelty slower than the Array method. This is something I will certainly consider in future.

As with a lot of things in Excel, this will come down to how the function is to be used in the real application. But for me one of the things I always appreciate about this site is that you never cease to learn something new on a pretty regular basis.

Thanks,
Steve.

0

We value your feedback.

Take our survey and automatically be enter to win anyone of the following:
Yeti Cooler, Amazon eGift Card, and Movie eGift Card!

I can only assume it's because Match is a pure Excel function, whereas Find isn't. (It may also help that Match has less functionality than Find().)

I suspect that from now on, my decision rules will be...
- Use the Cell method for non-expert OP's or when the table is small enough for acceptable response times.
- Use the Array method if the Cell method is too slow.
- Use the Match method if the Array method is too slow.

You got excellent advice on how to improve your function; there is not really much I could add of value.

However, I would advise you to not try and do this in a UDF, as what you are trying to do can be handled using an array formula, as described in my article here:

For example, let's say you have a list in A1:E1000. The column you want to check for a match is C:C, the column with the value to return is B:B, and you want the value corresponding to the third match.

As an array formula, you would enter it without typing the curly braces, and use Ctrl+Shift+Enter to finish it rather than just Enter. Excel will then display the curly braces to indicate the array status.

If you wanted, say, the value corresponding to the second-to-last match, then use LARGE():

And just so you know: this approach can be extended to cover multiple conditions, such as "fetch me the third value for where Col A = Smith and Col B = whatever". The article shows how to do that.

Thanks, Last_Free_Man, especially for the question!

0

Featured Post

ZipGrep is a utility that can list and search zip (.war, .ear, .jar, etc) archives for text patterns, without the need to extract the archive's contents.

One of a set of tools we're offering as a way to say thank you for being a part of the community.

Workbook link problems after copying tabs to a new workbook?
David Miller (dlmille)
Intro
Have you either copied sheets to a new workbook, and after having saved and opened that workbook, you find that there are links back to the original sou…

Excel can be a tricky bit of software to get your head around. Whilst you’ll be able to eventually get to grips with the basic understanding of how to get by, there are a few Excel tips that not everybody will even know about let alone know how to d…

The viewer will learn how to use the =DISCRINV command to create a discrete random variable, use this command to model a set of probabilities and outcomes in a Monte Carlo simulation, and learn how to find the standard deviation of a set of probabil…

The viewer will learn how to use a discrete random variable to simulate the return on an investment over a period of years, create a Monte Carlo simulation using the discrete random variable, and create a graph to represent the possible returns over…